swin-large-mask2former-finetuned-ER-Mito-LD8

This model is a fine-tuned version of facebook/mask2former-swin-large-ade-semantic on the Dnq2025/Mask2former_Pretrain dataset. It achieves the following results on the evaluation set:

• Mean Iou: 0.5775
• Loss: 34.1770

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0004
• train_batch_size: 4
• eval_batch_size: 4
• seed: 1337
• optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: polynomial
• training_steps: 6450

Training results

Training Loss	Epoch	Step	Mean Iou	Validation Loss
50.2838	1.0	129	0.3320	40.2146
38.6838	2.0	258	0.4777	33.3521
35.7266	3.0	387	0.5179	30.3853
28.8023	4.0	516	0.5172	31.1722
27.499	5.0	645	0.5405	29.2993
26.443	6.0	774	0.6063	27.7216
24.4565	7.0	903	0.5585	27.0999
23.7098	8.0	1032	0.5561	27.6352
22.5123	9.0	1161	0.5488	26.2882
21.6224	10.0	1290	0.5360	28.5500
21.1611	11.0	1419	0.5107	26.7347
20.0678	12.0	1548	0.5705	25.9258
19.8926	13.0	1677	0.5776	26.3917
18.5645	14.0	1806	0.5857	25.5181
18.504	15.0	1935	0.5701	26.0597
17.6968	16.0	2064	0.6108	25.1246
17.555	17.0	2193	0.6117	25.6074
17.2567	18.0	2322	0.5683	27.1555
16.0851	19.0	2451	0.6045	27.6046
16.308	20.0	2580	0.5550	28.1746
15.7719	21.0	2709	0.5898	25.3221
15.0966	22.0	2838	0.6299	27.0200
15.2529	23.0	2967	0.5870	29.0526
15.2963	24.0	3096	0.5638	27.0797
14.5228	25.0	3225	0.6203	27.8585
14.2121	26.0	3354	0.5659	28.6089
13.909	27.0	3483	0.6042	28.4436
14.0334	28.0	3612	0.6067	29.2367
13.3485	29.0	3741	0.5800	28.8674
13.4275	30.0	3870	0.6036	28.3902
13.1812	31.0	3999	0.5837	30.3429
13.0124	32.0	4128	0.5837	28.6284
12.4116	33.0	4257	0.5851	30.4995
13.3998	34.0	4386	0.5749	31.7624
12.794	35.0	4515	0.5840	29.0796
12.2829	36.0	4644	0.5594	30.8203
12.204	37.0	4773	0.6036	28.8408
12.6922	38.0	4902	0.5848	30.4332
12.1068	39.0	5031	0.5988	29.9606
11.7072	40.0	5160	0.5681	31.9938
11.7888	41.0	5289	0.5834	30.7203
11.5609	42.0	5418	0.5843	30.2104
11.4152	43.0	5547	0.6122	31.5076
11.932	44.0	5676	0.6020	31.8253
11.3475	45.0	5805	0.5828	32.4827
10.6893	46.0	5934	0.5792	33.6502
11.7356	47.0	6063	0.5777	33.7372
10.8846	48.0	6192	0.5716	34.2067
11.5715	49.0	6321	0.5709	34.1338
11.0337	50.0	6450	0.5775	34.0800

Framework versions

• Transformers 4.50.0.dev0
• Pytorch 2.4.1
• Datasets 3.3.2
• Tokenizers 0.21.0